avg_effects: Estimate average effects after CEA forests.

In bonander/CEAforests: Tree-based heterogenous cost-effectiveness and policy learning with individual-level data (beta)

Description

avg_effects Estimate de-biased average causal effects for the entire sample or for a specified subpopulation.

Usage

avg_effects(
  forest,
  WTP = NULL,
  subset = NULL,
  robust.se = FALSE,
  ci.level = 0.95,
  compliance.scores = NULL,
  icer.ci = TRUE,
  boot.ci = FALSE,
  R = 999
)

Arguments

forest	A trained CEA forest.
WTP	Willingness to pay per one-unit increase in Y.
subset	A specified subpopulation to obtain average effects for (optional).
robust.se	Whether or not robust (sandwich) standard errors are desired. Defaults to FALSE. Bootstrapped CIs or ICER CIs are not affected by this setting.
ci.level	The desired confidence level.
compliance.scores	An optional two-column matrix containing pre-fitted compliance scores for instrumental forests (col 1 = outcomes, col 2 = costs).
icer.ci	Logical, if confidence intervals for ICERs are desired. Uses Fieller's method if boot.ci=FALSE. Defaults to TRUE.
boot.ci	Logical, if bootstrapped confidence intervals (BCa) are desired. Defaults to FALSE.
R	Number of bootstrap replicates for bootstrapped CIs. Defaults to 999.

Value

Returns de-biased, pooled effect estimates with asymptotic variance estimates or accelerated percentile bootstrap confidence intervals if boot.ci=TRUE. Confidence intervals for ICERs are estimated using Fieller's method unless boot.ci=TRUE.

References

Chernozhukov, V., Chetverikov, D., Demirer, M., Duflo, E., Hansen, C., & Newey, W. (2017). Double/debiased/neyman machine learning of treatment effects. American Economic Review, 107(5), 261-65.

Examples

## Not run: 
To be added...

## End(Not run)

bonander/CEAforests documentation built on April 1, 2021, 10:57 a.m.